Skip to main content
Bitaboom is a TypeScript-first string utility toolkit focused on Arabic and bilingual (Arabic ↔ English) publishing workflows. This guide will walk you through the most common use cases with real code examples.

​
Installation

1

Install Bitaboom

Choose your preferred package manager:
npm install bitaboom
2

Import the functions you need

Bitaboom exports are tree-shakeable, so you only bundle what you use:
import { 
  makeDiacriticInsensitiveRegex, 
  removeMarkdownFormatting,
  preformatArabicText,
  estimateTokenCount,
  LLMProvider
} from 'bitaboom';

​
Core Features

​
Diacritic-Insensitive Matching

Build regular expressions that ignore Arabic diacritics, tatweel, and whitespace variants for flexible text search: 
import { makeDiacriticInsensitiveRegex } from 'bitaboom';

// Create a regex that matches regardless of diacritics
const pattern = makeDiacriticInsensitiveRegex('أنا إلى الآفاق');

// Matches text without diacritics
pattern.test('انا الي الافاق'); // true

// Also handles tatweel (Arabic text elongation)
pattern.test('أنــــا إلــــى الآفــــاق'); // true
This is perfect for search functionality in Arabic text applications where users might not type diacritics consistently.

​
Remove Markdown Formatting

Strip markdown syntax from text while preserving the actual content: 
import { removeMarkdownFormatting } from 'bitaboom';

const markdown = '**Bold** _italic_ [link](https://example.com)';
const plain = removeMarkdownFormatting(markdown);

console.log(plain); // "Bold italic link"
Common use cases:
  • Extracting plain text from markdown documents
  • Preparing text for search indexing
  • Generating preview snippets

​
High-Performance Arabic Preformatting

The preformatArabicText function is a single-pass optimizer for cleaning messy Arabic text, especially from OCR or scanned documents: 
import { preformatArabicText } from 'bitaboom';

const messy = 'بِسْمِ  اللَّهِ ( الرَّحْمَنِ ) 127 / 11 قَالَ ...';
const clean = preformatArabicText(messy);

// Fixes:
// - Multiple spaces → single space
// - Brackets spacing
// - Reference numbers
// - Ellipsis formatting
The preformatter consolidates 30+ formatting operations into a single optimized pass. It handles spacing, punctuation, brackets, ellipses, references, and more.

​
LLM Token Estimation

Accurately estimate token counts for Arabic text across different LLM providers. Arabic text uses approximately 3x more tokens than English due to BPE tokenization: 
import { estimateTokenCount } from 'bitaboom';

const text = 'بسم الله الرحمن الرحيم';
const tokens = estimateTokenCount(text);

console.log(`Estimated tokens: ${tokens}`);
Provider Efficiency Rankings:
  • Gemini: Most efficient (~25% fewer tokens than OpenAI)
  • OpenAI/Grok: Standard BPE baseline
  • Claude: Least efficient for Arabic (uses ~20% more tokens)

​
Common Workflow Patterns

​
Clean and Search Arabic Text

Combine preprocessing with search for robust Arabic text matching: 
import { 
  preformatArabicText, 
  makeDiacriticInsensitiveRegex 
} from 'bitaboom';

// Step 1: Clean the source text
const rawText = 'بِسْمِ  اللَّهِ ( الرَّحْمَنِ  الرَّحِيمِ ) ...';
const cleanText = preformatArabicText(rawText);

// Step 2: Create flexible search pattern
const searchPattern = makeDiacriticInsensitiveRegex('بسم الله');

// Step 3: Search in cleaned text
if (searchPattern.test(cleanText)) {
  console.log('Found match!');
}

​
Prepare Text for LLM Processing

Estimate tokens before sending to APIs to optimize costs and stay within limits: 
import { 
  preformatArabicText, 
  estimateTokenCount,
  LLMProvider 
} from 'bitaboom';

// Clean the text first
const rawArabic = 'نص عربي طويل جدا ...';
const cleaned = preformatArabicText(rawArabic);

// Estimate tokens for your provider
const tokens = estimateTokenCount(cleaned, LLMProvider.OpenAI);

// Check against model limits
const MAX_TOKENS = 4096;
if (tokens > MAX_TOKENS) {
  console.warn(`Text exceeds limit: ${tokens}/${MAX_TOKENS} tokens`);
  // Implement chunking strategy
}

​
Bilingual Document Processing

Handle mixed Arabic and English content: 
import { 
  removeMarkdownFormatting,
  getArabicScore,
  preformatArabicText 
} from 'bitaboom';

const bilingualText = `
**بسم الله** - _In the name of Allah_
قَالَ رَسُولُ اللَّهِ (ﷺ) - The Prophet said (peace be upon him)
`;

// Remove markdown formatting
const plain = removeMarkdownFormatting(bilingualText);

// Detect if text is primarily Arabic
const arabicRatio = getArabicScore(plain);
if (arabicRatio > 0.5) {
  // Apply Arabic-specific formatting
  const formatted = preformatArabicText(plain);
  console.log(formatted);
}

​
Interactive Demo

Explore all 80+ functions with live examples at bitaboom.surge.sh

​
Next Steps

Arabic Utilities

Diacritic handling, Urdu conversion, token estimation, and Arabic-specific text processing

Formatting & Typography

Smart quotes, punctuation spacing, bracket normalization, and 30+ typography helpers

Sanitization

Remove references, URLs, markdown, escaped spaces, and other noise from text

Transliteration

Normalize Arabic prefixes, clean up apostrophes, and replace salutations with ﷺ

Build docs developers (and LLMs) love

Get started for freeTalk to us